Abstract
We study the asymptotic Plateau problem in $\mathbb{H}_2\times \mathbb{R}$. We give the first examples of non-fillable finite curves with no thin tail in the asymptotic cylinder. Furthermore, we study the fillability question for infinite curves in the asymptotic boundary.
Introduction: The Asymptotic Plateau Problem
The Plateau problem is a classic question in geometric analysis: given a boundary curve, does there exist a minimal surface—a surface that locally minimizes area, like a soap film—that spans it? In the product space $\mathbb{H}^2 \times \mathbb{R}$ (the hyperbolic plane crossed with the real line), this problem is studied at the "asymptotic boundary," which topologically resembles a solid cylinder.
Historically, researchers identified a specific geometric obstruction known as a "thin tail"—a narrow, highly restricted portion of a curve—that prevents a curve from bounding a minimal surface. For a long time, it was conjectured that the absence of such a thin tail might be sufficient to guarantee fillability. This paper provides a definitive "No" to that conjecture.
The Generalized Butterfly Curve
The author constructs "Generalized Butterfly Curves" to demonstrate that even curves without thin tails can be non-fillable. These curves are "fat" and well-behaved in a local sense, yet they fail to bound a minimal surface due to a global mismatch between their height and width.
As shown in Figure 1, the curve $\Gamma$ consists of several segments ($\gamma^\pm, \beta^\pm$). If the central "neck" of the butterfly is squeezed too tightly relative to the overall height $M$, the geometry of the space prevents any minimal surface from spanning the gap. This highlights that fillability is not just a local property but depends on the proximity and relative scale of different curve components.
Methodology: Cages and Comparison Surfaces
To prove that these curves are non-fillable, the author employs a sophisticated "trapping" technique. The proof proceeds by contradiction: assuming a minimal surface exists, the author constructs a "cage" (a bounded region $\Omega_M$) using known minimal planes.
Once the hypothetical surface is confined within this cage (Figure 2), the author introduces a family of vertical catenoids—rotationally symmetric minimal surfaces. By applying the Maximum Principle, the author shows that as these catenoids are moved toward the cage, they would eventually touch the hypothetical surface in a way that violates the fundamental properties of minimal surfaces. This contradiction proves that no such spanning surface can exist.
Infinite Curves and the Corner Constraint
The paper also extends these investigations to infinite curves that reach the "caps" of the asymptotic cylinder at $\pm \infty$. The author identifies a new necessary condition: the curve must be nonoverlapping at the corner. If a curve overlaps with the circles where the cylinder walls meet the top or bottom caps, it cannot be filled. Furthermore, for an infinite curve to be fillable, its components at the caps must be geodesics.
Conclusion and Future Directions
This work shifts the understanding of the Plateau problem in $\mathbb{H}^2 \times \mathbb{R}$ by showing that "thin tails" are not the only obstruction to fillability. While the paper provides specific examples of non-fillable curves, a complete classification remains an open challenge. The author proposes a new conjecture: "tall" infinite curves should be fillable if and only if they are "fat at infinity," meaning they do not pinch down to a single point at the cylinder caps. This research underscores the extreme sensitivity of minimal surfaces to the global height and configuration of their boundaries in hyperbolic product spaces.
Summary updated